The global automotive infotainment market is valued at $35.78 billion in 2025, growing toward $37.95 billion in 2026 at a 7.19% CAGR. The OTA (over-the-air) vehicle software update market reached $5.2 billion in 2025 and is projected at $25 billion by 2035. Over 70% of new vehicles shipped globally now have OTA update capability.
These numbers describe a structural shift that predates both figures: the vehicle has become a software product. The competitive differentiation between vehicles at similar price points is increasingly measured in software roadmap velocity, HMI quality, and post-sale feature delivery — not in powertrain specifications. Engineering teams that understand how to build, ship, and update vehicle software at scale are the constraint.
Android Automotive OS: What 90% Adoption Actually Means
Android Automotive OS (AAOS) is projected to power nearly 90% of mid-to-high vehicle models shipped in 2026. The adoption list includes General Motors, Volvo, Ford, Honda, Renault, Stellantis, Hyundai/Kia, BMW, Volkswagen Group, Polestar, Lucid, Rivian, and BYD. The AAOS market itself is valued at $4.2 billion in 2025, growing toward $18.7 billion by 2034.
The critical nuance: only 6% of new vehicles shipping with AAOS in 2025 carry Google Automotive Services (GAS). OEMs want the OS without the Google ecosystem. The reasons are strategic (long-term vendor lock-in), commercial (Google’s compatibility requirements constrain customization), and competitive (OEM-developed digital services are higher-margin than Google’s).
This creates a specific engineering environment. AAOS provides the OS substrate — Linux kernel, Android framework, HIDL/AIDL HAL interfaces — but the infotainment application layer, digital services, and over-the-air update infrastructure are built by OEM software teams or Tier 1 suppliers. The vehicle software engineers are not writing Android apps — they are operating an Android-based platform and building proprietary applications on top of it.
Volvo’s January 2025 OTA deployment illustrates the production scale: the company pushed its largest-ever software update to 2.5 million vehicles globally at once, delivering a new “Volvo Car UX” with improved infotainment layouts, simplified visual hierarchies, and updated Google Maps integration. A single deployment decision reached 2.5 million devices simultaneously — with no dealer visit, no service appointment, no customer action required.
HMI Engineering: The Design Constraints That Are Regulatory
Human-Machine Interface design in modern vehicles is subject to safety constraints that have no equivalent in consumer software. ISO 15005 defines ergonomic principles for driver-system interaction, establishing cognitive load limits and distraction thresholds. Euro NCAP’s 2026 protocols have made this concrete in commercial terms: HMI design now directly affects safety ratings.
The 2026 Euro NCAP changes allocate 5 of 100 overall safety points to HMI design quality. Driver-monitoring systems expanded from 2 to 25 points. The specific penalty trigger: essential vehicle functions — horn, windscreen wipers, indicators, emergency SOS, climate control — buried in nested touchscreen menus. Two or more menu levels to reach any safety-critical function reduces the score. Research the Euro NCAP methodology references directly: touchscreen interactions for non-infotainment functions average 5–40 seconds of eyes-off-road time. Research consistently shows two seconds is where crash risk elevates significantly.
This creates a design constraint that runs counter to the consumer software instinct toward minimalism and large-screen immersion. In vehicle HMI:
- Physical buttons or top-level touch access are required for safety-critical functions
- Menu depth for any function used while driving is regulated by distraction research
- Night mode, ambient lighting, and display brightness adjustment must be accessible without attention allocation
- Multi-display environments (instrument cluster + central display + head-up display) require coherent interaction state across surfaces that may have different update rates and rendering contexts
The Qt framework is the primary development environment for automotive embedded HMI. Qt powers billions of automotive devices across Android Automotive, QNX, automotive-grade Linux, and Integrity RTOS. QML (Qt Modeling Language) is the declarative UI language; Qt Design Studio provides visual UI development with live preview on embedded targets. Qt is a Genivi Alliance member alongside BMW Group, Honda, and Volvo — which means its APIs and tooling are built around the industry’s reference architecture, not adapted from general-purpose development.
The SDV Leadership Map: What OEMs Are Actually Shipping
Tesla defined the software-defined vehicle category and remains the benchmark. But the competitive landscape is no longer one OEM ahead of everyone else.
Rivian is the clearest example of engineering organizational design for SDV. CEO RJ Scaringe’s explicit framing: zonal networks replace traditional wiring harnesses; domain controllers give way to centralized compute; software ownership stays in-house. Rivian deployed adaptive drive beam headlights before most incumbents by eliminating the traditional OEM-supplier dynamic where hardware specification constraints were driven by software capability limits — they designed the software capability first and spec’d the hardware to match.
Lucid shipped 13 OTA updates in 2025, with over 95% of Gravity features improvable post-sale. In H1 2027, Lucid will launch tiered ADAS subscriptions under DreamDrive Pro, priced $69–$199/month depending on autonomy level. This is recurring revenue from fixed hardware — a SaaS model applied to a physical vehicle. The engineering architecture that makes it possible is the same as any software subscription system: feature flags gated by entitlement checks, license validation integrated with the vehicle’s software management system, and OTA delivery of capability unlocks rather than physical hardware changes.
Mercedes-Benz MB.OS is the most architecturally ambitious traditional OEM effort: a purpose-built operating system with full cross-domain access across infotainment, automated driving, body & comfort, and charging. MB.OS is built on NVIDIA Orin system-on-chips with Luminar LiDAR for Level 3 autonomous capability. Post-2025, certain MB.OS functionalities are upgradable through the Mercedes me Store — in-vehicle commerce for software features.
Gartner’s Digital Automaker Index ranks software-defined vehicle capability: Tesla leads, followed by Chinese EV makers Nio and Xpeng, then Rivian and Lucid as the US EV challengers, with traditional OEMs — including Mercedes — trailing on software velocity despite investment scale.
Multimodal Interaction: Voice, Gesture, and LLM Integration
Voice user interfaces in vehicles are no longer voice-only. The 2025–2026 production standard is multimodal: LLM-backed voice assistants paired with ambient lighting feedback, haptic vibration response, head-up display confirmation, and gesture recognition — multiple output channels for a single user command.
The global voice assistant market in vehicles will surpass $30 billion by 2026. EV platforms represent over 40% of new multimodal development contracts in 2025–2026 across Asia-Pacific.
Hyundai’s 2026 LLM-based assistant (co-developed with Korean search giant Naver) represents the production frontier for in-vehicle NLP. At least 8 major Asia-Pacific OEMs announced NLP upgrades for production vehicles between 2025 and 2027. The shift from intent-classification voice systems (“navigate to [address]”) to LLM-backed conversation interfaces (“find a restaurant near the next highway exit that has parking for an EV”) requires fundamentally different backend architecture — context management, multi-turn conversation state, integration with vehicle sensor data, and low-latency inference that can respond within the human conversation window.
CarPlay and the OEM Control Question
Apple CarPlay Ultra (launched with the 2025 Aston Martin DBX) extends CarPlay into vehicle systems — climate control, radio — beyond the previous infotainment-only scope. The technical integration is more invasive than standard CarPlay, requiring AAOS or embedded system hooks into vehicle domain controllers.
OEM resistance is real: multiple manufacturers describe second-generation CarPlay Ultra as too intrusive for their integration comfort. GM’s strategic decision to phase out both CarPlay and Android Auto from all new vehicles beginning in 2025 is the clearest statement of OEM intent — they want to control the in-vehicle software relationship with the customer, not cede it to platform companies.
Android Auto’s 2025 update added Gemini AI integration (replacing Google Assistant), video streaming for parked vehicles, and web browsing — capabilities that expand its footprint beyond navigation and media into general-purpose computing when stopped.
Connected Services and the Subscription Monetization Gap
The connected vehicle services market was $21 billion in 2024, growing at 10.8% CAGR through 2034. OEMs are building subscription infrastructure for: ADAS feature unlocks, enhanced navigation services, in-car commerce (parking, tolls, EV charging payment), remote diagnostics, and fleet management APIs.
The gap in adoption: the percentage of consumers willing to pay for connected vehicle services dropped from 86% in 2024 to 68% in 2025 — an 18-point decline in one year. Safety, vehicle maintenance, and entertainment scored highest in consumer interest. Subscription features that consumers understand as clearly valuable (remote diagnostics, predictive maintenance alerts, ADAS capability unlocks) have a much clearer conversion path than features that feel like capabilities that should have shipped with the vehicle by default.
The engineering requirement for sustainable connected services: tight feature-to-subscription mapping, in-vehicle purchase flows that do not distract from driving, and OTA delivery of capability unlocks that activates within minutes of purchase. The backend is standard subscription SaaS architecture; the challenge is the in-vehicle UX of the purchase and activation flow.
How we approach this at Insoftex
Automotive software product design requires a different engineering discipline than consumer app development — not because the technologies are exotic, but because the regulatory context, the safety constraints, and the multi-domain integration requirements create failure modes that consumer software engineers do not encounter.
For clients building HMI systems or in-vehicle software products, we design against the distraction research and Euro NCAP scoring criteria from the beginning — not as a compliance retrofit. A UI designed with physical-control fallbacks and two-tap maximum depth for any in-motion function is a better UI by the metric that matters: it does not distract the driver.
For OEMs or Tier 1 suppliers building connected service platforms and OTA infrastructure, we design the subscription model and feature-entitlement architecture in the first architecture session — because the deployment architecture for a feature (optional unlock vs. standard with paid upgrade) determines the update delivery model, the license validation approach, and the vehicle software build configuration.
Building vehicle software, HMI systems, or connected service platforms? Our automotive engineering team works on connected vehicles, diagnostics, telemetry, and HMI applications. Start with a Product Pilot for architecture review, regulatory constraint mapping, and working prototype in three weeks.